WY14643 Increases Herpesvirus Replication and Inhibits IFNβ Production Independently of PPARα Expression

PPAR agonists are used clinically to treat both metabolic and inflammatory disorders. Because viruses are known to rewire host metabolism to their own benefit, the intersection of immunity, metabolism, and virology is an important research area. ABSTRACT Peroxisome proliferator activated receptor (PPAR) agonists are commonly used to treat metabolic disorders in humans because they regulate fatty acid oxidation and cholesterol metabolism. In addition to their roles in controlling metabolism, PPAR agonists also regulate inflammation and are immunosuppressive in models of autoimmunity. We aimed to test whether activation of PPARα with clinically relevant ligands could impact gammaherpesvirus infection using murine gammaherpesvirus-68 (MHV68, MuHV-4). We found that PPAR agonists WY14643 and fenofibrate increased herpesvirus replication in vitro. In vivo, WY14643 increased viral replication and caused lethality in mice. Unexpectedly, these effects proved independent of PPARα. We found that WY14643 suppressed production of type I interferon after MHV68 infection in vitro and in vivo. Taken together, our data indicate that caution should be employed when using PPARα agonists in immuno-metabolic studies, as they can have off-target effects on viral replication through the inhibition of type I interferon production. IMPORTANCE PPAR agonists are used clinically to treat both metabolic and inflammatory disorders. Because viruses are known to rewire host metabolism to their own benefit, the intersection of immunity, metabolism, and virology is an important research area. Our article is an important contribution to this field for two reasons. First, it shows a role for PPARα agonists in altering virus replication. Second, it shows that PPARα agonists can affect virus replication in a manner independent of their predicted target. This knowledge is valuable for anyone seeking to use PPARα agonists as a research tool.

also orchestrate fatty acid oxidation, particularly in muscle (8). Consistent with their roles in cellular fatty acid metabolism, activation of PPARs by agonists such as fibrates (PPARa) and thiazolidinediones (PPARg ) has demonstrated clinical efficacy in treating metabolic disorders in humans (9)(10)(11). Importantly, fatty acid metabolism has been shown to regulate host-virus interaction and contribute to determining the outcome of infection (12,13).
Moreover, PPARs regulate inflammation, and PPAR agonists are used clinically to reduce inflammation in atherosclerosis, diabetes, neurodegenerative diseases, and autoimmune diseases (10). Working through diverse pathways, PPAR agonists repress NF-k B and AP-1 DNA binding, regulate nitric oxide production, inhibit dendritic cell maturation, reduce cytokine expression by effector T cells, and inhibit leukocyte recruitment to sites of inflammation, all of which are known key regulators of cellular antiviral response (14,15). Collectively, this evidence highlights the potential roles of PPARs in shaping immune responses against DNA viruses.
Conversely, it is known that herpesviruses manipulate host cell metabolism during infection to promote viral replication and chronic infection (16,17), including through the induction of peroxisomes (18,19). A recent report found that HCMV induces peroxisome biogenesis to enhance plasmalogen synthesis, which is required for efficient HCMV envelopment (19). Herpesviruses also encode viral proteins that target peroxisomes, suggesting that modulation of peroxisomal function is important for these viruses (20)(21)(22).
Despite their immunoregulatory functions, our understanding of the effects of PPAR agonists on infectious disease outcomes remains incomplete. There are contradictory reports suggesting that synthetic agonists or dietary lipids improve or impair resistance to pathogen challenge, and the molecular mechanisms of PPAR-mediated immunoregulation during infection remain elusive (23)(24)(25)(26)(27).
Thus, we sought to test the hypothesis that PPAR agonists could alter host-herpesvirus interactions. We evaluated agonists of PPARa, PPARb/d , and PPARg and found that the compounds WY14643 and fenofibrate (agonists of PPARa) produce strong proviral effects. Treatment with these compounds increased the replication of herpesviruses in two different taxa. Consistent with our hypothesis, this effect initially seemed dependent on PPARa expression. However, anomalous data led us to carefully control for mouse genetic background and microbiome and revealed that these proviral effects occur independently of PPARa. We further demonstrate that WY14643 treatment in macrophages reduced interferon (IFN)-b and interferon stimulated gene (ISG) expression after virus infection. WY14643's proviral and interferon-suppressing effects were seen in vivo as well, significantly increasing MHV68 replication and animal mortality and decreasing ISG transcription.

RESULTS
PPARa agonists promote herpesvirus replication in vitro. To examine the effects of PPAR activation on DNA virus infection, we used murine gammaherpesvirus-68 (MHV68) as our model. MHV68 readily infects mice and undergoes phases of infection similar to human gamma-herpesviruses such as Kaposi's sarcoma-associated herpesvirus (KSHV) and Epstein Barr virus (EBV) (28,29). The replication of MHV68 was measured in bone marrow-derived macrophages (BMDMs), as this virus infects and replicates in macrophages (as well as B cells and dendritic cells) in vivo (28,30).
The effects of PPAR activation on MHV68 replication were examined by treating BMDMs with agonists for PPARa (fenofibrate or WY14643), PPARb/d (GW501516), or PPARg (rosiglitazone) prior to infection. After infection, we quantified viral replication with flow cytometry, staining for the expression of lytic viral proteins on the surfaces of infected cells (31). We found that PPARa agonists fenofibrate and WY14643 both increased expression of lytic viral proteins on infected macrophages (Fig. 1A). However, GW501516 and rosiglitazone had no effect on MHV68 replication, indicating that PPARb/d and PPARg agonists do not regulate MHV68 replication in BMDMs at the doses commonly used in the literature to stimulate PPARb/d and PPARg (Fig. 1A). We confirmed these effects of WY14643 and fenofibrate with viral growth curves at high and low high multiplicity of infection (MOI), both of which showed increased MHV68 replication in macrophages treated with fenofibrate or WY14643 compared to untreated cells (Fig. 1B, C). Moreover, effects of PPARa agonists fenofibrate and WY14643 increased with increasing dose (Fig. 1D, E) Thus, treatment with PPARa agonists increases MHV68 replication.
To determine if PPARa agonists affect replication of other herpesviruses that infect BMDMs, we tested whether WY14643 or fenofibrate would increase replication of murine cytomegalovirus (MCMV), a betaherpesvirus that also readily infects BMDMs. We found that replication of MCMV was also increased by these treatments (Fig. 1F), indicating that these effects of PPARa agonists apply to at least one other subfamily of herpesvirus.
We wondered if PPARa agonist effects could promote virus replication even if cells were treated with agonist after infection, or if the effects of agonist required pretreatment. To test this, we compared three different treatment protocols. We pretreated, as before, with PPARa agonists overnight and replaced agonist in the media following infection with MHV68 (pre/post). We compared this with pretreatment only (pre) or post treatment only (post). We found that pretreatment with PPARa agonists, fenofibrate, and WY14643 was required to increase MHV68 replication (Fig. 2). We found no increase in virus replication when cells were treated postinfection with agonist. Growth curves of MHV68 in macrophages isolated from C57BL/6J mice after pretreatment with vehicle control (DMSO), WY14643 or fenofibrate. Cells were infected with MHV68 at MOI = 5 (B) or MOI = 0.1 (C). Virus was quantitated by plaque assay on 3T12 cells. The data represent the mean 6 SD from 5 independent experiments. D-E. Growth curves of MHV68 in macrophages isolated from C57BL/6J mice with different doses of WY14643 (D) or fenofibrate (E). Virus was quantitated by plaque assay on 3T12 cells. The data represent the mean 6 SD from 3 independent experiments. (F) Growth curves of MCMV in BMDMs isolated from C57BL/6J mice. After pretreatment with vehicle, WY14643, or fenofibrate, cells were infected with MCMV at MOI = 1. The data represent the mean 6 SD from 2 independent experiments. FACS data were found to be normal and analyzed with one-way ANOVA and Tukey's multiplecomparison tests. Data are all shown as mean 6 SD; *, P , 0.05; **, P , 0.01; ***, P , 0.001; ****, P , 0.0001.
We also found that pretreatment alone was sufficient to increase virus replication. These data suggest that PPARa agonists alter the cellular environment prior to infection in such a way that enhances virus replication.
WY14643 increases virus replication independently of expression of PPARa. We next questioned whether the effects of PPARa agonists WY14643 and fenofibrate depend on expression of PPARa. To answer this question, we initially performed viral growth curve experiments in BMDMs isolated from C57BL/6J or Ppara 2/2 mice obtained from Jackson Laboratories (here denoted "Ppara 2/2 /J"). In these experiments, the fenofibrate and WY14643 increased viral replication (Fig. 3A). These effects were abolished in the Ppara 2/2 /J BMDMs (Fig. 3B). This suggested that their proviral activity involves their canonical role as PPARa agonists. However, we noted an increase in virus replication in vehicle-treated Ppara 2/2 /J macrophages relative to control macrophages from C57BL/6J mice (Fig. 3B), which was unexpected.
Questioning whether this was due to differences in mouse genetic background or microbiome differences between knockout and C57BL/6J mice, we generated littermatecontrolled mice by crossing C57BL/6J mice with Ppara 2/2 /J mice to obtain heterozygous offspring. These heterozygous mice were interbred to yield both knockout ("Ppara 2/2 ") and wild-type ("Ppara 1/1 ") animals that were used for BMDM generation. When we compared viral replication in these littermate-controlled macrophages, we observed no baseline difference in virus replication between the genotypes in the vehicle group (Fig. 3C, D). However, the effects of fenofibrate and WY14643 remained intact in the littermatecontrolled knockout BMDMs; agonist treatment was associated with increased virus replication in PPARa-deficient cells just as in wild-type cells (Fig. 3D). This suggests that fenofibrate and WY14643 increase virus replication independently of PPARa expression.
At first, our in vitro data suggested that the proviral effects of fenofibrate and WY14643 depend on PPARa. However, this seems to be an artifact of different mouse genetic backgrounds or microbiome differences between the C57BL/6J and knockout colonies; when we used littermate controls, this dependency vanished. This indicates that WY14643 increases MHV68 replication independently of its canonical role as agonists of PPARa.
WY14643 suppresses the interferon response by reducing type I IFN production. We next asked whether the increase in MHV68 replication we observed with WY14643 required the type I interferon receptor, which is important for controlling MHV68 replication (32,33). We first tested this question with a genetic knockout experiment, isolating BMDMs from Ifnar 2/2 and C57BL/6J mice. The cells were treated as before with WY14643, fenofibrate, or a control and then infected with MHV68 at a low MOI (MOI = 0.1), such that increases in virus replication due to WY14643 treatment could

Increased Herpesvirus Replication by PPARa Agonists
Microbiology Spectrum be observed during the multistep growth curve. As expected, WY14643 and fenofibrate increased virus replication in wild-type cells (Fig. 4A). MHV68 also predictably replicated to a higher level in Ifnar 2/2 cells compared to the wild-type cells (Fig. 4B). In contrast to wild-type cells, agonist treatment did not further increase virus replication in Ifnar 2/2 BMDMs (Fig. 4B), suggesting that WY14643 and fenofibrate effects depend on type I interferon receptor signaling.
To confirm these results, we measured viral replication after blocking IFNAR in C57BL/6J BMDMs with an anti-IFNAR antibody. At 72 hours postinfection, WY14643 treatment increased virus replication compared to the isotype control. Consistent with the results in Fig. 5A and 5B, treatment with the anti-IFNAR antibody increased viral replication compared to the isotype control. However, combining WY14643 with the anti-IFNAR antibody did not cause an additional increase in virus replication (Fig. 4C), indicating that WY14643-mediated increase in MHV68 replication requires IFNAR.
To clarify the role of type 1 interferon in agonist-driven phenotypes, we used RT-qPCR to examine Ifnb and ISG expression. We quantified Ifnb, Isg20, Isg15, and Cxcl10 transcripts in BMDMs from C57BL/6J mice infected with MHV68 for 6 hours. We saw that MHV68 infection induced the transcription of these genes, and that WY14643 treatment significantly attenuated this effect (Fig. 4D to G). We also noted that WY14643 treatment significantly reduced the transcripts of these genes compared to the control even in the absence of viral infection (Fig. 4D to G).
Repeating these experiments in cells from littermate-controlled Ppara 1/1 and Ppara 2/2 mice, we confirmed that the effect of WY14643 on Ifnb, Isg15, and Isg20 is PPARa-independent ( Fig. 4H to K). Notably, we again observed that WY14643 reduced the transcripts of Ifnb, Isg15, Isg20, and Cxcl10 without infection, though the effect was not significant for Isg20 in Ppara 2/2 cells (Fig. 4H to K). This suggests that WY14643 suppression of Ifnb gene and ISG expression is independent of PPARa expression.
WY14643 increases MHV68 replication and lethality independent to PPAR-a expression. Because we observed that WY14643 increased viral replication and decreased IFNb production in vitro, we tested whether WY14643 increased viral replication in mice. First, we used C57BL/6J and Ppara 2/2 /J mice to test the effects of in vitro treatment with WY14643. Mice were injected with WY14643 or a vehicle control for 7 days, starting 3 days prior to infection and continuing for 4 days after infection (Fig. 5A). We chose to pretreat mice with WY14643 by intraperitoneal injection because of our in vitro data indicating that pretreatment was necessary (Fig. 2). Using luciferase-tagged MHV68 (MHV68-M3FL), we infected mice intraperitoneally and imaged them over multiple days to measure acute . Viral growth was measured over 96 h by plaque assay on 3T12 cells. (C) Virus replication (MOI = 0.1) in C57BL/67 BMDMs treated with WY14643 and anti-IFNAR antibody. Cells were pretreated with DMSO or WY14643 along with 5 mg/mL anti-IFNAR antibody (clone MAR1-5A3, Biolegend number 127312) or an IgG isotype control, then infected. Cells were collected at 0 h and 72 h postinfection. Viral growth was measured by plaque assay on 3T12 cells, and data are shown as the geometric mean 6 geometric SD of 8 independent experiments. (D-G) BMDMs from C57BL/6J mice were pretreated with vehicle or WY14643 for 16 h prior to infection with MHV68. Then, RT-qPCR was performed to quantify transcripts of Ifnb (n = 8), Isg20 (n = 7), Isg15 (n = 6), and Cxcl10 (n = 5) before and 6 h after MHV68 infection. Relative expression of these genes is shown normalized to Gapdh. (H-K) BMDMs from littermate-controlled Ppara 1/1 or Ppara 2/2 mice were pretreated with vehicle or WY14643 for 16 h prior to infection with MHV68. Then, RT-qPCR was performed to quantify transcripts of Ifnb (n = 5), Isg20 (n = 6), Isg15 (n = 7), and Cxcl10 (n = 6) 6 h after MHV68 infection. Relative expression of these genes is shown normalized to Gapdh. (C-K) Data are all shown as the mean 6 SD; *, P , 0.05; **, P , 0.01; ***, P , 0.001; ****, P , 0.0001. The distribution of RT-qPCR data were checked and found to follow a normal distribution. Plaque assay data were found to follow a lognormal distribution and log-transformed before analysis with one-way ANOVA and Tukey's multiple-comparison test. RNA transcripts from vehicle and agonist-treated cells were compared with one-tailed paired t tests.
virus replication (31,34). Although it is not physiological, we chose the intraperitoneal route of infection for MHV68 because it is well characterized to lead to equivalent levels of viral replication and latency compared with intranasal (35). Additionally, we chose intraperitoneal infection because we also used intraperitoneal injection to administer WY14643. We found that C57BL/6J mice treated with WY14643 had increased virus replication compared to vehicle-treated mice (Fig. 5B)   Data represent 2 independent experiments. Transcripts from peritoneal exudate cells were compared with one-tailed unpaired t tests. Data all shown as mean 6 SD; *, P , 0.05, **, P , 0.01, ***, P , 0.001, ****, P , 0.0001. In vivo imaging data were found to follow a lognormal distribution. They were analyzed with 2-way repeated measures ANOVA tests and Tukey's multiple-comparison tests after log-transformation. The graphed data are untransformed. Survival data were analyzed with Mantel-Cox tests.

Increased Herpesvirus Replication by PPARa Agonists
Microbiology Spectrum absent in Ppara 2/2 /J mice treated with WY14643 (Fig. 5C). Surprisingly, even though mice were infected with a dose of MHV68 that does not cause lethality in wild-type mice, C57BL/6J mice treated with WY14643 succumbed to infection (Fig. 5D) at a frequency similar to mice deficient in the type I interferon receptor (32,33,36). This lethality effect was not present in the PPARa knockout mice (Fig. 5D). As with our in vitro data, these in vivo data give the appearance that WY14643 increases lethality and virus replication through its canonical function as an agonist of PPARa. However, after repeating these experiments with mice interbred in our colony, we no longer observed the PPARa-dependency of WY14643. Littermate-controlled knockout mice treated with WY14643 succumbed to virus infection at a similar rate to wild type littermate controls (Fig. 5E) and also displayed significantly increased virus replication at day 7 (Fig. 5F, G).
To determine whether WY14643 treatment in vivo suppressed induction of interferon stimulated genes after MHV68 infection, we pretreated mice with WY14643 or control and infected them with MHV68 intraperitoneally as in Fig. 5A. Two days after infection, we sacrificed the mice and collected peritoneal cells by lavage to measure ISG expression. We found that expression of Isg20 and Cxcl10 was decreased in WY14643-treated mice (Fig. 5H). These data support the hypothesis that WY14643 suppressed interferon responses in vivo, and that this may contribute to increased replication and lethality with MHV68 infection.

DISCUSSION
We determined that two agonists of PPARa, WY14643 and fenofibrate, increase herpesvirus replication independently of expression of PPARa. In macrophages, WY14643 and fenofibrate increased replication of both a gamma-and beta-herpesviruses. We determined that WY14643 suppressed type I interferon induction after MHV68 infection. In vivo, WY14643 increased MHV68 replication and lethality in infected mice, a strong effect given that MHV68 is rarely fatal. The increase in replication and lethality was not dependent on expression of PPARa, but this phenomenon was only revealed when we used littermate control mice for comparison. When we measured ISG expression in peritoneal cells after intraperitoneal infection, we observed decreased ISG expression. Although we cannot conclude the relative contribution of impaired type I interferon to WY14643-induced increase in MHV68 replication and lethality, these data suggest that WY14643 suppressed type I interferon in vivo. Given the importance of type I interferon in controlling acute MHV68 infection, we propose that the inhibition of interferon may be one contributing factor to the lethality phenotype. Importantly, the effects of WY14643 on herpesvirus infection and interferon response are independent of PPARa.
Our results imply that WY14643 increased MHV68 replication in part by decreasing interferon production, but further work will be required to describe a more specific mechanism. We observed that WY14643 treatment reduced the transcription of interferon and interferon-stimulated genes even without virus infection. This was true in both wild type and littermate control Ppara 2/2 BMDMs. These data suggest that WY14643 regulates basal interferon expression. We do not think this represents global downregulation of transcription by WY14643, both because of previous reports of transcriptomic analysis after agonist treatments and from our own analysis of global gene expression (data not shown) (37).
Although we did not find a PPARa-dependent function of agonists on MHV68 replication, we have not ruled out the possibility that PPARs play a role in virus replication. Such interactions remain worth investigating. One mechanism by which PPARs could regulate virus replication is through the upregulation of negative regulators of inflammation such as Ik B and the soluble IL-1 receptor antagonist (38). A second mechanism is by regulating inflammatory gene expression directly. PPARs decrease NF-k B and AP-1 activities through transrepression, which stabilizes corepressor complexes on inflammatory gene promoters, such as nitric oxide synthetase, IL-1b, and IL-12 (39,40).
Our results indicate that researchers should be mindful of PPAR-independent immune modulation caused by PPAR agonists. Separating PPARa-dependent and independent effects may be complicated by genetic modifiers and microbiome differences in mouse models, as was the case in our experiments. Initially, we observed a PPARa-dependent phenotype when comparing Ppara 2/2 /J and C57BL/6J macrophages. The effects of WY14643 and fenofibrate on virus replication were abolished in Ppara 2/2 /J cells, suggesting that their mechanism is through their canonical role as PPARa agonists. However, MHV68 replication was unexpectedly high in vehicle-treated Ppara 2/2 /J macrophages compared with the control, leading us to question the validity of the comparison. This anomalous virus replication, along with the illusion of PPARa-dependency, disappeared when we generated and used littermate-controlled Ppara 1/1 and Ppara 2/2 mice. There are several possible reasons for this. For one, the Jackson Laboratory does report that the Ppara 2/2 /J mice have 3 single-nucleotide polymorphism markers that are still of the 129S4/SvJae allele-type, which could contribute to differences in immune response and virus replication. Additionally, microbiome differences between our C57BL/6J and Ppara 2/2 /J colonies, which were normalized by interbreeding the two, could have played a role. These results are similar to other reports describing the off-target effects of etomoxir, an inhibitor commonly used to block Cpt1a and fatty acid oxidation (41)(42)(43). Our results highlight the importance of using littermate controls when trying to establish genetic dependency.

MATERIALS AND METHODS
Animals. C57BL/6J, B6;129S4-Ppara tm1Gonz /J (44), and B6.129S2-Ifnar1 tm1Agt /Mmjax (45), were purchased from The Jackson Laboratory. All mice were housed under specific pathogen-free, double-barrier facility at the University of Texas Southwestern Medical Center. Mice were fed autoclaved rodent feed and water. Knockout strains were maintained by breeding homozygous knockout males with homozygous knockout females, producing litters with only the knockout genotype.
For isolation of BMDMs, male mice were used. For in vivo imaging and survival curves, a mixture of male and female mice was used. Mice were maintained and used under a protocol approved by UT Southwestern Medical Center Institutional Animal Care and Use Committee (IACUC).
Virus infection. Fully differentiated BMDMs were seeded on 24-well plates (1.5 Â 10 5 cells per well) or 6-well plate (10 6 cells per well). Cells were pretreated with either vehicle control (0.1% DMSO) or agonists. Fenofibrate was used at a final concentration 50 mM and WY14643 was used at a final concentration of 200 mM (48) Rosiglitazone was used at a final concentration of 1 mM (49) and GW501516 at 100 nM (50) for 16 hours. The next day, macrophages were infected with MHV68 at multiplicity of infection (MOI) = 5 or 0.1. For MCMV experiments, cells were infected at MOI = 1. After an hour, cells were washed with PBS twice to remove unabsorbed viruses. Then, culture medium containing treatments was added to the wells. For growth curves, samples were collected at 0 h, 24 h, 48 h, 72 h and 96 h after infection and were frozen at 280°C. The titer of virus was determined by plaque assay in 3T12 cells. For flow cytometric analysis, cells were collected 24 h after infection.
For IFNAR-blocking experiments, mature BMDMs were pretreated overnight for 16 hours with anti-IFNAR antibody (5 mg/mL) or isotype control and WY14643 (200 mM) or DMSO control. Cells were infected with MHV68 at MOI = 0.1. After 1 hour, the cells were washed twice with PBS to remove unabsorbed virus. Culture media with the same antibody and drug treatments was added to cells. Samples were taken at 0 h and 72 h postinfection and frozen at 280°C. The titer of virus was determined by plaque assay in 3T12 cells.
Plaque assay. The concentration of virus was measured by plaque assay in 3T12 cells. The frozen samples containing viruses were thawed in an incubator. The samples were serially diluted, then added to a monolayer of 3T12 cells. After an hour of absorption, the cells were then covered with 1% methylcellulose. Plates were incubated at 37°C for 7 days, and the plaques were stained with 0.1% crystal violet.
RT-qPCR. BMDMs in 6-well plates were either infected with MHV68 at MOI = 5 for 6 hours or treated with DMXAA (10 m/mL) for 4 hours. RNA was extracted using Qiagen RNeasy minikit (Qiagen) and reverse transcribed into cDNA using SuperScript VILO cDNA Synthesis kit (Thermo Fisher Scientific). Relative quantification of target genes was determined using PowerUp SYBR green Master Mix (Thermo Fisher Scientific) in a QuantStudio 7 Flex real-time PCR system. Primers used for amplifying target genes are listed in Table 1.
MHV68 acute replication in mice. Experiments were carried out using 8-12 weeks old mice under the protocol approved by IACUC. Mice were injected intraperitoneally with either vehicle control (15% HS15 in normal saline) or WY14643 (100 mg/kg) for 1 week starting from 3 days before virus infection. Mice were then infected with MHV68-M3FL at the dose of 10 6 PFU through intraperitoneal route (31,34). To quantify virus-encoded luciferase expression, mice were weighed and injected with 150 mg/kg of d-Luciferin (GOLDBIO) immediately prior to imaging using an IVIS Lumina III In Vivo Imaging System (PerkinElmer). A region of interest (ROI) was drawn around the abdominal region and applied to all mice. Total flux (photons/second) of the ROI was measured (exposure = 1 sec., F/stop = 1.2, FOV=E, bin-ning=medium, emissions filter=open) using Living Image software (PerkinElmer). Survival of the mice was recorded until 20 days after infection.
Quantification and statistical analysis. All data are presented as mean 6 SD. Statistical comparisons were performed using GraphPad Prism 9.4 software. In vitro gene transcripts were compared using paired one-tailed t tests. Ex vivo gene transcripts were compared using unpaired one-tailed t tests. In vivo imaging data were analyzed using one-way ANOVA with Tukey's multiple comparisons post-test. Statistical significance was set at P , 0.05. The numbers of independent replicates (n) are reported in the figure legends.